1. Field of the Invention
The present invention relates to an image pickup device in which pixels are arranged in a matrix and which generates pixel signals for image and pixel signals for focus detection, and an image pickup apparatus.
2. Description of the Related Art
In recent years, an image pickup device in which pixels are arranged in a matrix and which generates pixel signals for image and pixel signals for focus detection has been proposed. Such an image pickup device includes, for example, a configuration in which pixels for focus detection are discretely arranged among pixels for image arranged in a matrix, and a configuration in which a plurality of photo diodes are provided within one pixel, so that both pixel signals for image and pixel signals for focus detection are obtained from one pixel.
Further, while a high-resolution image is desired for a still image, an image with fixed resolution (fixed number of pixels) with a frame rate of, for example, 30 fps or 60 fps is desired to be acquired for a movie. Therefore, upon acquisition of a movie, or the like, the number of pixels to be read out is reduced through pixel addition, so as to reduce time required for readout and improve a frame rate.
Also in an image pickup device having pixels for focus detection, a technique of performing pixel addition has been proposed.
For example, Japanese Patent Application Laid-Open Publication No. 2010-20055 discloses an image pickup apparatus having an addition readout mode in which output of a plurality of pixels within an image pickup device is added and read out, and if a group of pixels to be added includes only pixels for image pickup, the pixels are added, while if the group of pixels to be added includes pixels for focus detection, signals of a group of pixels for image pickup and signals of pixels for focus detection are read out so as not to be mixed with each other.
Further, Japanese Patent Application Laid-Open Publication No. 2013-110607 discloses a technique of, in an image pickup device including a pixel having a micro lens, a plurality of photoelectric conversion regions disposed for each pixel, and addition means for adding signals in the photoelectric conversion regions disposed for each pixel, setting an addition readout region in which a plurality of pixels are arranged in two dimensions of a horizontal direction and a vertical direction and which outputs a signal obtained by adding the signals of the photoelectric conversion regions out of the plurality of pixels by the addition means, and an independent readout region which outputs a signal without addition of the signals of the photoelectric conversion regions by the addition means, and making the number of pixels in a horizontal direction of the addition readout region identical with that of the independent readout region.
Pixel addition which has been conventionally performed will be described with reference to FIG. 5 and FIG. 7 to FIG. 9 according to embodiments of the present invention.
First, it is assumed that, for example, as shown in FIG. 7, a configuration of a pixel unit 21 in which pixels having photo-diodes are arranged in a matrix is basically a primary color Bayer array (here, diagonal hatching from upper left to lower right indicates green pixels, vertical hatching indicates red pixels, and horizontal hatching indicates blue pixels), and pixels for focus detection R, L, T and B are discretely arranged in the primary color Bayer array. Further, as shown in FIG. 5, an image pickup device includes an analog memory unit 22 which respectively stores analog pixel signals read out from the pixel unit 21, pixel signals are read out for each row from the memory unit 22, and a plurality of pixel signals within one row are converted into digital signals simultaneously at a column-parallel AD converter 25.
FIG. 8 shows an example of V2/2 H2/2 addition, and FIG. 9 shows an example of V1/5 H3/3 addition. Here, pixel addition is performed so that pixel signals of the same color are added, and V indicates a vertical direction, H indicates a horizontal direction, a denominator indicates for each of how many pixels of the same color pixels in an array are added, and a numerator indicates the number of pixels to be added.
In the case of V2/2 H2/2 addition shown in FIG. 8, pixel signals of the same color of all the pixels are added for each of 2×2 pixels in the vertical direction and in the horizontal direction, and converted into digital signals by the column-parallel AD converter 25.
In the case of V1/5 H3/3 addition shown in FIG. 9, pixel signals of one in every five rows are converted into digital signals by the column-parallel AD converter 25, and three pixels of pixel signals of the same color converted into the digital signals are added in the horizontal direction.
By the way, the number of times of operation of the column-parallel AD converter 25 largely affects a frame rate of an image, and generally, time required for reading out an image is proportional to the number of times of operation of the column-parallel AD converter 25.
Therefore, while, in the case of V2/2 H2/2 addition, it requires readout time of the same level as that in the case of readout of all the pixels, in the case of V1/5 H3/3 addition, because it requires readout time of approximately ⅕ of that in the case of readout of all the pixels or the V2/2 H2/2 addition, it is possible to increase a frame rate by substantially 5 times.